@Firaxis & 2k - Good AI & Netcode is KEY to an enjoyable Beyond Earth experience!!

[Ryoga]

if Firaxis and 2k had paid even a bit of attention to this forum before (and they did) then they already know that a lot people complains about Civ V's AI. we've been doing it for the past... 4 years at the very least.

They did try to make some adjustments to that with G&K and BNW and even a few patches. It got better, but it didn't really change that much.


OP, it's not just that it is too late to change BE at this point, you are late to the party of AI complaints as well.
It was said, they know, they tried to fix it as much as they could, this is the best that we got with Civ V, and as for the best that we'll get with BE, we'll see soon enough.

 

[Magma_Dragoon]

The AI's use of workers isn't so much the problem as how they just don't build any. Civ5's AI is fairly good when it's flavor settings aren't suicidal. And I don't think savable build queues would be useful in BE since it has the same expansion limiting elements as civ 5.

 

[Acken]

The AI's use of workers isn't so much the problem as how they just don't build any.

What ? The AI makes workers, it even start with one on Emperor+.

 

[rawrkitty]

So i took the time to set up a VM and load up civnet to make a point, not that any of you guys care, but call it catharsis

So here we see, in the first two images, we can set auto build to our cities to basically become puppets with priorities as denoted by those icons





Now things get more interesting, we can create custom build orders that priortize and check buildings and adhere to the order, i've had 30+ ques before, it's wonderful.



And our final screen shows us it in action, when we set a city to auto build mode - custom, it will automatically execute the build order, allowing us to save precious time to actually play the game instead of micromanaging every little thing.
in 1-10 cities, micromanaging each element fine. at 50 cities, it makes tedium outscale fun.

This kind of system is crucial to multiplayer and reducing time.

Especially in civ5 where "clicking" a single button (such as adding something to a que) requires a response from the host, and then sends it after response is recieved and then requires a response to reflect the pushed packets with the changed data, you often notice 5-10 second delays between "clicking" something and it acutally happening, this means players often spend 5+ minutes each turn just clicking buildings in their cities. - easily mitigated with out fixing any netcode

None of this is dumbing down the game, or automating it and playing it for you, it's just having governors follow your orders.

You are literally doing the same thing you would otherwise be doing, but with essentially a macro.
These are choices that we as players should have :l

 

[Magma_Dragoon]

What ? The AI makes workers, it even start with one on Emperor+.

Sometimes it builds lots of workers and has several large cities will all the tiles improved and it conquers the continent, builds all the wonders, has a huge military, and keeps tech parity with a human player on immortal.

Sometimes it refuses to build workers, it has one or two free from difficulty level and one from liberty and they spend more time running back and forth between tiles than building improvements.

Most of the time does something in between where it has not nearly enough workers and flounders in the mid-game because it spent so much time working unimproved tiles in the early game.

I had one game where Russia had 5 cities and 2 workers. They bounced from one end of Russia to the other, improving one tile in one city, then spend ten turns walking to the next tile to be improved. Moscow was size 12 at turn 150.

The next game I play, Russia had 8 cities with 13 workers. She built several wonders, had top military score, was neck and neck with me in literacy, and posed a huge threat to me. Moscow was size 27 at turn 150.

It's very disappointing to see the AI fail because it rolled a low tile improvement flavor, or a low growth flavor.

 

[KrikkitTwo]

Sometimes it builds lots of workers and has several large cities will all the tiles improved and it conquers the continent, builds all the wonders, has a huge military, and keeps tech parity with a human player on immortal.

Sometimes it refuses to build workers, it has one or two free from difficulty level and one from liberty and they spend more time running back and forth between tiles than building improvements.

Most of the time does something in between where it has not nearly enough workers and flounders in the mid-game because it spent so much time working unimproved tiles in the early game.

I had one game where Russia had 5 cities and 2 workers. They bounced from one end of Russia to the other, improving one tile in one city, then spend ten turns walking to the next tile to be improved. Moscow was size 12 at turn 150.

The next game I play, Russia had 8 cities with 13 workers. She built several wonders, had top military score, was neck and neck with me in literacy, and posed a huge threat to me. Moscow was size 27 at turn 150.

It's very disappointing to see the AI fail because it rolled a low tile improvement flavor, or a low growth flavor.

Definitely.. those flavors should be moderated so they don't influence game behavior into stupid ways.


And I think that player automation for both city buildings and tile improvements is necessary

 

[Galgus]

What is the point of this rant? Do you think they can rewrite the netcode or the AI from scratch in a month because you told them to? Either the things you're talking about have already been fixed (which I doubt), or there's no way they can be in the short time left before release.

I've seen mods that at least improved the AI: I have some hope that the actual devs addressed the #1 complaint of Civ 5.

 

[joncnunn]

They did try to make some adjustments to that with G&K and BNW and even a few patches. It got better, but it didn't really change that much.

It was said, they know, they tried to fix it as much as they could, this is the best that we got with Civ V, and as for the best that we'll get with BE, we'll see soon enough.

Yup, both expansions were already locked into the flavor driven system. In addition an inspection of the source code made available to Civ V modders shows it's a mess and it's actually amazing that the developers managed to make any improvements to the AI in G&K and BNW.

The AI's use of workers isn't so much the problem as how they just don't build any. Civ5's AI is fairly good when it's flavor settings aren't suicidal. And I don't think savable build queues would be useful in BE since it has the same expansion limiting elements as civ 5.

Flavor settings are indeed suicidal with the wrong position. A human knows if meets Zulu on turn 10 to start preparing immediately. The building or wonder flavors waits until AI Zulu already has his army built up which is too late.

Meanwhile, a human would know by turn 30 if there's no one on their landmass and so not need to build large armies. Agressive AI flavors would build the big army anyway.

And actually, the AI's "ideal worker to city" ratio is actually 1.5 workers to 1 city in Civ V. While that was standard for Civ IV, there's actually higher than most humans in Civ V. It is notoriously slow to connect spare luxuries compared to a human, which basically causes AIs in low unique resource diversification happiness issues even when it knows several other AIs.

What ? The AI makes workers, it even start with one on Emperor+.

No, that's Immortal+ in which the Civ V AI starts with a worker. Emperor is the first level in which the AI starts with a Scout in addition to the warrior.

I've seen mods that at least improved the AI: I have some hope that the actual devs addressed the #1 complaint of Civ 5.

That's basically the "Smart AI" mod and some other mods that start with that. They help on the margins.

 

[DarkestOnion]

I am not the biggest fan of 'Automating Workers' or 'City Governors' in Civilization games because I feel that delegating those choices can lead to suboptimal play; particularly at the highest levels of play. As such, I am not a fan of either encouraging or further enabling/designing towards what I think of as a 'lazy play style'. Of course, given the negative conotation of the word lazy it may be important to note that I am judging the lazy actions as a strategy, not casting personal aspersions on the character of any player who chooses such lazy strategies (sic this is not an ad hominem attack or a trollbait paragraph).

That being said the AI could use some work in the Civilization series imho. However, as I am not an AI Programmer at Firaxis, I must admit that I do not fully understand the specific challenges involved in crafting a superior AI to the one presented in Civ 5. As a result, I find myself reluctant to get onboard with any suggestions to improve the AI offered by other non-AI Programmers on either the CFC or Civ:BE Steam Forums. Put simply, I know that I do not know enough about AI Programming to know if what I might like to see in the field of AI Improvement in Civ:BE is either realistic or even possible with the engine being used.

I understand your frustration with an AI that is not enough of a challenge OP. I also understand how tempting it is to recommend a Copy/Paste of AI Programming from a game you prefer (Gal Civ 2 for example) into Civ:BE. From what little I know of AI Programming (from second hand knowledge from a game developer friend of mine admittedly) it is both illegal to Copy/Paste AI code from another game and liekly not to work due to Game Engine incompatibilities. Therefore, I fear that any desire for AI Improvements must be channeled through AI Programmers. Sadly, I have never heard a peep from 2K/Firaxis AI Programmers on CFC, so I think this post will amount to little more than hopeful speculation.

 

[Technopinky]

Now this doesn't seem unreasonable, but when you have a 50 city empire and you're in the modern age and you have to create a 30 building que each time you found a new city, when 99% of the time you're using the SAME PRIORITY ANYWAY, there's no reason for the feature that USED to exist to not exist anymore.

Well you're in luck! Because from everything we've seen about CivBE you're only going to have about 5-6 cities in the late game.

 

[AriochIV]

i'm more than a little disappointed to see the kind of hostility towards these things, i guess i'll just have to suffer through them and live with it.

I want radically better AI every bit as much as you do, but we're not going to get it in this product. If you expect it, you're going to be disappointed.

I've seen mods that at least improved the AI: I have some hope that the actual devs addressed the #1 complaint of Civ 5.

I have no such hope. According to Dennis Shirk, the AI in Civ V BNW is in "an amazing place."

 

[Thormodr]

The AI in Civilization 5 was poor and I expect the same for Civilization 5:BE.

Maybe they'll start from scratch for Civ VI and make a competent AI.

In reference to the Op, I like what Stardock has been doing lately. Firaxis could learn a thing or two from them.

 

[rastak]

The AI in Civilization 5 was poor and I expect the same for Civilization 5:BE.

Maybe they'll start from scratch for Civ VI and make a competent AI.

In reference to the Op, I like what Stardock has been doing lately. Firaxis could learn a thing or two from them.

Not a helluva lot from my experience which goes all the way back to Sid's F15 Strike Eagle.

And loading up OS/2 Warp to play Gal Civ......the first.

 

[Seek]

Well you're in luck! Because from everything we've seen about CivBE you're only going to have about 5-6 cities in the late game.

I don't know where you've been looking, but everything I've seen points to ICS on an unprecedented scale.

So it will probably be somewhere in the middle.

 

[Thormodr]

Not a helluva lot from my experience which goes all the way back to Sid's F15 Strike Eagle.

And loading up OS/2 Warp to play Gal Civ......the first.

Gal Civ III looks intriguing. 64 bit opens up a whole new world of possibilities.

For example, a 100 player game.

 

[proftc]

Gal Civ III looks intriguing. 64 bit opens up a whole new world of possibilities.

For example, a 100 player game.

Hmm... 100 players at full graphics... and then the flames start coming out of the video card.

 

[Thormodr]

Hmm... 100 players at full graphics... and then the flames start coming out of the video card.

Manageable, if programmed correctly.

Anyway, there were no meltdowns reported.

 

[nguajajaja]

i feel the urge to bring this thread back to the surface and hope the devs see this.

At least make the AI not buggy like in Civ 5, or make the AI easy to mod so that the community can enhance its intelligence.

 

[Karl0413]

Manageable, if programmed correctly.

Anyway, there were no meltdowns reported.

That's because they burned up any eyewitnesses in said meltdowns.

 

[Lord Shadow]

Sure, some of rawrkitty's complaints are valid. I would never oppose a better AI, but it's unrealistic to expect something revolutionary from Beyond Earth. As for the automation, frankly, I like it where it is. Firstly, you no longer have to maintain 30+ cities, and secondly, you don't apply the exact same build order to every city you found anymore. Sure, there are some common things, but they're neither extensive nor tedious to deal with.

As for GalCiv2, well, I haven't played in ages, but as a game it has it a lot easier: no geography, no movement restrictions, and barely any "terrain" improvements which aren't a no-brainer. Easy expansion, tech race, build ships and attack, or shift planetary development looking towards the other victory types: it doesn't have much in the way of challenges for a decent AI.

And regarding praise to Stardock... I shouldn't have to bring up Elemental, and the fact its two expansions haven't really elevated it to anything more than a playable game. Rather bland and generic, Legendary Heroes could never hold me for long. It's funny how some despite Firaxis' decent track record dread CivBE, but after Stardock's trainwreck and blandness are most optimistic about GalCiv3. GalCiv2 was the last really good thing they did, but that was 8 years ago and had its fair share of issues for a while.